Temperature gradients across the dies of today's high performance very large scale integration (VLSI) components, such as a microprocessor, can adversely affect component performance. For example, a temperature variation between two clock driver circuits within a microprocessor often results in a skew in the system clock of the microprocessor. Moreover, the die of the microprocessor may reach an unacceptable temperature that causes the microprocessor to malfunction or stop functioning.
To protect a microprocessor from thermal damage, a diode is typically placed in the die of the microprocessor to provide a die temperature indication. This diode is driven with a fixed amount of current, and the corresponding voltage drop across the diode provides an indication of the microprocessor temperature. Unfortunately, the diode provides a temperature indication that is accurate to about ±10° C., which is often not accurate enough to provide an early indication of a temperature abnormality. Moreover, because a single diode is typically utilized to monitor the die temperature of the entire microprocessor, only a single temperature indication is provided.
Given the size and complexity of integrated circuits, such as microprocessors, it is extremely difficult to determine and monitor a temperature gradient across the integrated circuit using only a single diode positioned at a location in the die of the integrated circuit. As such, substantial variations in temperature across the die of the integrated circuit can go undetected. Consequently, early indications that a thermal problem exists in a portion of the integrated circuit go undetected.